Case studies

Thirteen cloud, DevOps and security engagements across healthcare, SaaS, fintech, ERP and education. Clients are described by sector and region only, the same way we would protect yours. Named references are available on request, with each client's consent.

Healthcare and health payments

01Healthcare · US · AzureDelivered

Secure partner file exchange on Azure

A US value services organisation for physician practices

The challenge

  • Partners must send files from known people and networks only
  • Dev and QA domains lacked edge security and certificates
  • No current architecture map or view of Azure spend

What we delivered

Secure file exchange
SFTP on Azure Storage: a container per partner, key-only, IP-locked
Edge security
Cloudflare proxy, DNS and managed certificates for dev and QA
Central secrets
Environment keys moved into a vault for dev, staging and prod
Architecture and cost
A live architecture document and a costed Azure savings report
Outcome

Each partner exchanges data through its own key-only, IP-restricted endpoint, and the client has a live architecture reference and a costed list of Azure savings.

02Healthcare claims · US · AWSDelivered

Production releases a claims platform can trust

A US third-party administrator processing healthcare claims

The challenge

  • Production releases depended on one engineer's laptop
  • Production and QA shared one server and one service
  • Cloud keys sat in plain text in production config

What we delivered

Release ownership
Production deploys for the .NET API and the React front end
Risk discovery
Showed a routine QA push could restart the live claims service
Secrets remediation
Key rotation and a managed secrets store, planned and documented
Full runbook
Repos, deploys, checks, logs and troubleshooting per environment
Outcome

Releases are repeatable and documented, and the client can see and close its two biggest risks: shared environments and exposed keys.

03Healthcare · Global · AWSDesigned, awaiting go-ahead

Ten security findings, turned into a plan

A global medical and telehealth provider for crew and passengers at sea

The challenge

  • Production database publicly reachable for 11 days
  • Unencrypted database traffic and long-lived server keys
  • Manual releases and an unannounced database engine upgrade

What we delivered

Remediation plan
An owner, a fix and a date for all ten findings, with live traffic kept safe
Private network design
Database behind a bastion, with a keyless Session Manager option
Identity design
Scoped instance roles replace static keys; UAT split from production
Release discipline
Pipeline plan with approval gates and UAT tests for upgrades
Outcome

Every finding has an owner, a fix and a date, and the client reviewed the network design before anything was built, as it had asked.

04Behavioural health · US · AWSDelivered

Automated releases for a behavioural health EHR

A US multi-clinic psychiatry and TMS therapy group

The challenge

  • Releases of its own EHR were not fully automated
  • New background jobs needed a queue in UAT and production
  • A legal-entity change had to leave live systems untouched

What we delivered

Pipeline automation
End-to-end release automation, set out with the lead developer
Queue design
SQS for UAT and production, with environments kept separate
Account governance
Two-step AWS legal-entity change with no impact on workloads
Working alongside
Advised the client's developers so their team owns the result
Outcome

The team has an agreed path to automated EHR releases and a production-ready queue design, with clinical systems untouched by the account change.

05Virtual care · US · AWSDelivered

Safe releases for on-demand virtual care

A US urgent-care group connecting patients to clinicians by live video

The challenge

  • Every release reaches patients in the middle of care
  • Release knowledge sat with one engineer
  • Staff needed production access without shared logins

What we delivered

Production releases
Deployments run jointly with the client's own developers
Named access
Individual IAM users for client staff, no shared credentials
Knowledge transfer
A step-by-step walkthrough of the production release for staff
Repeatable process
One documented release path the whole team follows
Outcome

Releases follow a known, repeatable process, and the client deploys with its own named, auditable accounts instead of shared logins.

06Healthcare payments · US · AWSDelivered

A payments worker that scales with its queue

A US payments company serving dental practices

The challenge

  • Payment jobs had to survive restarts with nothing lost
  • Uneven load: a fixed fleet wastes money or falls behind
  • The pipeline had to reach AWS without stored keys

What we delivered

Queue worker on Fargate
Deletes a job only on success; drains in-flight work on deploy
Failure handling
Four retries, then a dead-letter queue with an instant alarm
Backlog auto-scaling
1–4 workers in production; QA and UAT scale to zero when idle
Keyless CI/CD
Terraform for three environments; GitLab CI with OIDC into AWS
Outcome

The worker absorbs spikes on its own, test environments cost nothing when idle, and no cloud keys are stored in the pipeline.

Beyond healthcare

07Hospitality SaaS · US · AWSDelivered

SOC 2 in-house and a leaner AWS bill

A US loyalty platform for multi-location restaurant and hospitality brands

The challenge

  • Paying for a compliance SaaS with nobody to run SOC 2
  • An AWS bill of about $4,300 a month for around 10 customers
  • Production nodes averaging around 5% CPU

What we delivered

SOC 2 workspace
Evidence, controls, policies and risk inside the client's own app
Subscription replaced
Compliance SaaS cancelled; SOC 2 pen-test plan written
Cost review
A read-only audit of every service, commitment and utilisation
Savings roadmap
$929 a month found without touching production data paths
Outcome

The compliance subscription is gone. A phased plan saves 22% of AWS spend now and about 57% once reserved commitments end in July 2027.

08Digital agency platform · US · AWS + AzureDesigned, awaiting go-ahead

Free, auto-renewing HTTPS for 218 websites

A US digital agency platform hosting its clients' websites

The challenge

  • 156 of 218 hosted domains had no working HTTPS
  • 62 sites relied on paid certificates renewed by hand
  • The edge proxy ran Ubuntu 14.04, unsupported since 2019

What we delivered

Domain audit
295 domains checked for DNS, certificate and expiry status
Edge redesign
nginx and Let's Encrypt replace HAProxy; certificates renew themselves
Zero-DNS cutover
Traffic moves by reassigning one IP; rollback is one command
Portal recovery
Client portal rebuilt from its build output and redeployed on Azure
Outcome

Once approved, every hosted site gets trusted, auto-renewing HTTPS, paid renewals stop, and clients who bring their own certificates keep them.

09Fintech · US · AWSDelivered

Release runbooks for a retail investing app

A US platform giving investors weekly ETF signals and one-click rebalancing

The challenge

  • Release steps lived with one engineer
  • Frontend production releases were done by hand
  • Release ownership had to move to the client's developers

What we delivered

Production operations
Ran the AWS hosting: nginx, PM2 processes and certificates
Runbook
Architecture, inventory, CI/CD, deploys and integrations
Scripted release
A production frontend deploy script handed to the dev team
Clean handover
Developers now release without depending on one person
Outcome

Releases are documented end to end, and the platform's developers ship frontend changes without waiting on infrastructure.

10Manufacturing web estate · US · SecurityDelivered

Closing the gaps in a web security assessment

An HVAC brand's public web estate, assessed through a reseller partner

The challenge

  • An earlier assessment had left parts of the estate unchecked
  • Sites spread across HubSpot, WP Engine and AWS hosting
  • wp-admin IP allow-listing on WP Engine was misbehaving

What we delivered

Reconnaissance
DNS reconnaissance across the domain and its subdomains
TLS findings
Weak cipher findings documented on HubSpot-hosted subdomains
API key remediation
Google Maps API key remediated on the AWS-hosted application
Credential exposure
Hardcoded credentials in wp-config.php flagged for rotation
Outcome

The client got a complete picture: TLS, key and credential issues with fixes, and the wp-admin access block traced to the custom theme's own files.

11SaaS · US · AWSIn progress

Finding the bottleneck and the savings

A platform running on AWS EC2 and Amazon RDS for SQL Server

The challenge

  • A production performance incident to get to the bottom of
  • A mandate to cut EC2 and RDS spend
  • Production and UAT sharing one single-AZ database

What we delivered

Incident investigation
Traced the slowdown to RDS SQL Server bottlenecks
Commitment analysis
Reserved Instances vs Savings Plans across EC2 and RDS
Resilience call
Multi-AZ proposed for the shared database, cost stated openly
Client-facing advice
Findings and options presented directly to the client
Outcome

The client knows where its performance problem sits, which commitments cut its bill, and why one recommendation adds cost to remove a single point of failure.

12ERP · US · AWSDesigned, awaiting go-ahead

Major upgrades for an eight-database ERP fleet

An ERP platform running on eight Amazon RDS for PostgreSQL databases

The challenge

  • Eight PostgreSQL 14 databases due for a major upgrade
  • Downtime had to stay short and controlled
  • Many application servers had to be updated in step

What we delivered

Upgrade runbook
PostgreSQL 14 to 17 using RDS Blue/Green Deployments
Maintenance page
A custom page shown to users during the upgrade window
Fleet scripts
Deployment scripts that update every server in one pass
Rollback ready
The old environment stays until the new one is validated
Outcome

The fleet has a documented path to PostgreSQL 17 with a short switchover, users see a clear maintenance page, and rollback stays available until sign-off.

13Education · India · Cloud migrationDesigned, awaiting go-ahead

Moving a legacy school ERP to the cloud

Two Indian schools sharing one on-premise ERP server

The challenge

  • A Visual FoxPro 9 ERP on one on-premise Windows Server
  • 5–10 staff working in it at once over RDP
  • Licensing rules that generic cloud quotes usually miss

What we delivered

Option comparison
AWS, Azure and Windows VPS hosts priced on the same spec
Licensing check
RDS CALs costed; lapsed Software Assurance flagged early
Right-sizing
Storage cut from 1 TB to 300 GB against about 88 GB in use
Recommendation
AWS Lightsail Windows, with a lower-cost VPS as the backup route
Outcome

The schools have three costed routes, with licensing risks surfaced before any move. Next: a discovery and compatibility audit.

Which of these looks like your platform?

We'll walk you through the closest case study and what we would do first.

Book a free cloud audit